Intelligent energy saving power system control of telecom DC power plant

This paper presents the design of intelligent energy saving power system controller (IESPSC) in a DC power plant of Korean Telecommunications. From past experience, rectifiers for telecom DC power plants have been very inefficient at light loads. The proposed control scheme offers a "novel load sharing" approach based on the knowledge of each unit's efficiency in parallel rectifiers. Neural networks are used to identify a unit's efficiency characteristic curve corresponding to load currents, which is in turn utilized to produce a system efficiency close to the maximum under all operating conditions. A dramatic saving in operation costs can be achieved by limiting the number of operating units to the minimum and maintaining high efficiency at the determined loads.